Prospective evaluation of interrater agreement between EEG technologists and neurophysiologists

AbstractWe aim to prospectively investigate, in a large and heterogeneous population, the electroencephalogram (EEG)-reading performances of EEG technologists. A total of 8 EEG technologists and 5 certified neurophysiologists independently analyzed 20-min EEG recordings. Interrater agreement (IRA) for predefined EEG pattern identification between EEG technologists and neurophysiologits was assessed using percentage of agreement (PA) and Gwet-AC1. Among 1528 EEG recordings, the PA [95% confidence interval] and interrater agreement (IRA, AC1) values were as follows: status epilepticus (SE) and seizures, 97% [96–98%], AC1 kappa = 0.97; interictal epileptiform discharges, 78% [76–80%], AC1 = 0.63; and conclusion dichotomized as “normal” versus “pathological”, 83.6% [82–86%], AC1 = 0.71. EEG technologists identified SE and seizures with 99% [98–99%] negative predictive value, whereas the positive predictive values (PPVs) were 48% [34–62%] and 35% [20–53%], respectively. The PPV for normal EEGs was 72% [68–76%]. SE and seizure detection were impaired in poorly cooperating patients (SE and seizures; p < 0.001), intubated and older patients (SE; p < 0.001), and confirmed epilepsy patients (seizures; p = 0.004). EEG technologists identified ictal features with few false negatives but high false positives, and identified normal EEGs with good PPV. The absence of ictal features reported by EEG technologists can be reassuring; however, EEG traces should be reviewed by neurophysiologists before taking action.

Download Full-text

Automated Inter-Ictal Epileptiform Discharge Detection from Routine EEG

Healthier Lives, Digitally Enabled - Studies in Health Technology and Informatics ◽

10.3233/shti210012 ◽

2021 ◽

Author(s):

Duong Nhu ◽

Mubeen Janmohamed ◽

Lubna Shakhatreh ◽

Ofer Gonen ◽

Patrick Kwan ◽

...

Keyword(s):

Royal Melbourne Hospital ◽

Epileptiform Discharges ◽

Interictal Epileptiform Discharges ◽

Model Variant ◽

Eeg Recordings ◽

Generalized Epilepsy ◽

Electroencephalogram Eeg ◽

Health Hospital ◽

Epilepsy Monitoring

Epilepsy is the most common neurological disorder. The diagnosis commonly requires manual visual electroencephalogram (EEG) analysis which is time-consuming. Deep learning has shown promising performance in detecting interictal epileptiform discharges (IED) and may improve the quality of epilepsy monitoring. However, most of the datasets in the literature are small (n≤100) and collected from single clinical centre, limiting the generalization across different devices and settings. To better automate IED detection, we cross-evaluated a Resnet architecture on 2 sets of routine EEG recordings from patients with idiopathic generalized epilepsy collected at the Alfred Health Hospital and Royal Melbourne Hospital (RMH). We split these EEG recordings into 2s windows with or without IED and evaluated different model variants in terms of how well they classified these windows. The results from our experiment showed that the architecture generalized well across different datasets with an AUC score of 0.894 (95% CI, 0.881–0.907) when trained on Alfred’s dataset and tested on RMH’s dataset, and 0.857 (95% CI, 0.847–0.867) vice versa. In addition, we compared our best model variant with Persyst and observed that the model was comparable.

Download Full-text

Zero-crossing patterns reveal subtle epileptiform discharges in the scalp EEG

Scientific Reports ◽

10.1038/s41598-021-83337-3 ◽

2021 ◽

Vol 11 (1) ◽

Author(s):

Jan Pyrzowski ◽

Jean- Eudes Le Douget ◽

Amal Fouad ◽

Mariusz Siemiński ◽

Joanna Jędrzejczak ◽

...

Keyword(s):

Normal Subjects ◽

Intracranial Eeg ◽

Zero Crossing ◽

Temporal Precision ◽

Scalp Eeg ◽

Epileptiform Discharges ◽

Interictal Epileptiform Discharges ◽

Eeg Recordings ◽

Patients With Epilepsy ◽

Low Amplitude

AbstractClinical diagnosis of epilepsy depends heavily on the detection of interictal epileptiform discharges (IEDs) from scalp electroencephalographic (EEG) signals, which by purely visual means is far from straightforward. Here, we introduce a simple signal analysis procedure based on scalp EEG zero-crossing patterns which can extract the spatiotemporal structure of scalp voltage fluctuations. We analyzed simultaneous scalp and intracranial EEG recordings from patients with pharmacoresistant temporal lobe epilepsy. Our data show that a large proportion of intracranial IEDs manifest only as subtle, low-amplitude waveforms below scalp EEG background and could, therefore, not be detected visually. We found that scalp zero-crossing patterns allow detection of these intracranial IEDs on a single-trial level with millisecond temporal precision and including some mesial temporal discharges that do not propagate to the neocortex. Applied to an independent dataset, our method discriminated accurately between patients with epilepsy and normal subjects, confirming its practical applicability.

Download Full-text

Quantitative Measure to Differentiate Wicket Spike from Interictal Epileptiform Discharges

Communication in Biomathematical Sciences ◽

10.5614/cbms.2021.4.1.2 ◽

2021 ◽

Vol 4 (1) ◽

pp. 14-22

Author(s):

Suryani Gunadharma ◽

Ahmad Rizal ◽

Rovina Ruslami ◽

Tri Hanggono Achmad ◽

See Siew Ju ◽

...

Keyword(s):

Discriminant Analysis ◽

Linear Discriminant Analysis ◽

Quantitative Measure ◽

Linear Discriminant ◽

Sharp Waves ◽

Epileptiform Discharges ◽

Interictal Epileptiform Discharges ◽

Temporal Location ◽

Eeg Recordings ◽

Excellent Accuracy

A number of benign EEG patterns are often misinterpreted as interictal epileptiform discharges (IEDs) because of their epileptiform appearances, one of them is wicket spike. Differentiating wicket spike from IEDs may help in preventing epilepsy misdiagnosis. The temporal location of IEDs and wicket spike were chosen from 143 EEG recordings. Amplitude, duration and angles were measured from the wave triangles and were used as the variables. In this study, linear discriminant analysis is used to create the formula to differentiate wicket spike from IEDs consisting spike and sharp waves. We obtained a formula with excellent accuracy. This study emphasizes the need for objective criteria to distinguish wicket spike from IEDs to avoid misreading of the EEG and misdiagnosis of epilepsy.

Download Full-text

A STUDY ON COMPARATIVE YIELDS OF STANDARD SHORT TERM ELECTROENCEPHALOGRAM AND LONG TERM ELECTROENCEPHALOGRAM RECORDING IN SUSPECTED EPILEPSY PATIENTS

Pakistan Armed Forces Medical Journal ◽

10.51253/pafmj.v71i5.5921 ◽

2021 ◽

Vol 71 (5) ◽

pp. 1727-31

Author(s):

Saima Shafait ◽

Wasim Alamgir ◽

Imran Ahmad ◽

Saeed Arif ◽

Jahanzeb Liaqat ◽

...

Keyword(s):

Diagnostic Yield ◽

Extended Period ◽

Final Analysis ◽

Epileptiform Discharge ◽

Military Hospital ◽

Recording Time ◽

Epileptiform Discharges ◽

Interictal Epileptiform Discharges ◽

Interictal Epileptiform Discharge ◽

Electroencephalogram Eeg

Objective: To compare the yield of interictal epileptiform discharges on prolonged (1-2 hours) electroencephalogram (EEG) as compared to standard routine (30 minutes) electroencephalogram (EEG). Study Design: Comparative observational study. Place and Duration of Study: Pak Emirates Military Hospital, Rawalpindi from Oct 2019 to Sep 2020. Methodology: A total of 364 outdoor patients with suspected epilepsy were recruited for the study. Out of these 55 electroencephalograms were excluded after applying exclusion criteria and 309 were included for final analysis. Electro-encephalograms were recorded using a 10-20 international system of electrode placement. The duration of each standard electroencephalogram was 30 minutes. It was followed by recording for an extended period of 60 minutes at least. The time to the appearance of the first abnormal interictal epileptiform discharge was noted. For analytical purposes, epileptiform discharges were classified as “early” if they appeared within the first 30 minutes and as “late” if appeared afterward. All electro-encephalograms were evaluated independently by two neurologists. Results: A total of 309 electroencephalograms were included for final analysis. Interictal epileptiform discharges were seen in 48 (15.6%) recordings. The mean time to appearance of first interictal epileptiform discharge was 14.6 ± 19.09 minutes. In 36 (11.7%) cases, discharges appeared early (within the first 30 minutes) whereas in the remaining 12 (3.9%) cases, discharges appeared late. This translates into a 33% increase in the diagnostic yield of electroencephalogram with an extended period of recording. Conclusion: Extending the electroencephalogram recording time results in a significantly better diagnostic yield of outdoor electroencephalogram.

Download Full-text

Incidence of Electrographic Seizures in Patients With COVID-19

Frontiers in Neurology ◽

10.3389/fneur.2021.614719 ◽

2021 ◽

Vol 12 ◽

Author(s):

Brandon L. Waters ◽

Andrew J. Michalak ◽

Danielle Brigham ◽

Kiran T. Thakur ◽

Amelia Boehme ◽

...

Keyword(s):

Critical Illness ◽

Case Series ◽

Nasopharyngeal Swab ◽

Convulsive Status Epilepticus ◽

Retrospective Case ◽

Predictive Values ◽

Electrographic Seizure ◽

Epileptiform Discharges ◽

Interictal Epileptiform Discharges ◽

Electrographic Seizures

Critical illness and sepsis are commonly associated with subclinical seizures. COVID-19 frequently causes severe critical illness, but the incidence of electrographic seizures in patients with COVID-19 has been reported to be low. This retrospective case series assessed the incidence of and risks for electrographic seizures in patients hospitalized with COVID-19 who underwent continuous video electroencephalography monitoring (cvEEG) between March 1st, 2020 and June 30th, 2020. One hundred and twenty-two patients were initially identified who resulted SARS-CoV-2 nasopharyngeal RT-PCR swab positivity with any electroencephalography order placed in the EMR. Seventy-nine patients met study inclusion criteria: age ≥18 years, >1 h of cvEEG monitoring, and positive SARS-CoV-2 nasopharyngeal swab PCR. Six (8%) of the 79 patients suffered electrographic seizures (ES), three of whom suffered non-convulsive status epilepticus. Acute hyperkinetic movements were the most common reason for cvEEG in patients with ES (84%). None of the patients undergoing cvEEG for persistent coma (29% of all patients) had ES. Focal slowing (67 vs. 10%), sporadic interictal epileptiform discharges (EDs; 33 vs. 6%), and periodic/rhythmic EDs (67 vs. 1%) were proportionally more frequent among patients with electrographic seizures than those without these seizures. While 15% of patients without ES had generalized periodic discharges (GPDs) with triphasic morphology on EEG, none of the patients with ES had this pattern. Further study is required to assess the predictive values of these risk factors on electrographic seizure incidence and subsequent outcomes.

Download Full-text

Amplitude-integrated EEG amplitudes differ between wakefulness and sleep states in children

10.1101/2021.07.29.21261253 ◽

2021 ◽

Author(s):

Verena Tamara Loeffelhardt ◽

Adela Della Marina ◽

Sandra Greve ◽

Hanna Mueller ◽

Ursula Felderhoff-Mueser ◽

...

Keyword(s):

Sensitivity Analysis ◽

Antiepileptic Drugs ◽

Deep Sleep ◽

Epileptiform Discharges ◽

Interictal Epileptiform Discharges ◽

Continuous Background ◽

Eeg Recordings ◽

Physiological Background ◽

Light Sleep

Introduction: Interpretation of pediatric amplitude-integrated EEG (aEEG) is hindered by the lack of knowledge on physiological background patterns in children. The aim of this study was to assess the amplitudes and bandwidths of background patterns during wakefulness and sleep in children from long-term EEGs. Methods: Forty long-term EEGs from patients < 18 years of age without or only solitary interictal epileptiform discharges were converted into aEEGs. Upper and lower amplitudes (μV) of the C3 - C4, P3 - P4, C3 - P3, C4 - P4, and Fp1 - Fp2 channels were measured during wakefulness and sleep. Bandwidths (BW, μV) were calculated, and sleep states assessed during the episodes of interest. A sensitivity analysis excluded patients who received antiepileptic drugs. Results: Median age was 9.9 years (interquartile range 6.1 - 14.7). All patients displayed continuous background patterns. Amplitudes and BW differed between wakefulness (C3 - C4 channel: upper 35 (27 - 49), lower 13 (10 - 19), BW 29 (21 - 34)) and sleep. During sleep, episodes with high amplitudes (upper 99 (71 - 125), lower 35 (25 - 44), BW 63 (44 - 81)) corresponded to sleep states N2 - N4. These episodes were interrupted by low amplitudes that were the dominating background pattern towards the morning (upper 39 (30 - 51), lower 16 (11 - 20), BW 23 (19 - 31), sleep states REM, N1, and N2). With increasing age, amplitudes and bandwidths declined. The sensitivity analysis yielded no differences in amplitude values or bandwidths. Conclusion: aEEG amplitudes and bandwidths were low during wakefulness and light sleep and high during deep sleep in stable children undergoing 24 hour EEG recordings. aEEG values were not altered by antiepileptic drugs in this study.

Download Full-text

Automatic Detection of Interictal Epileptiform Discharges Based on the Similarity of Time-Series Sequence

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.239-240.921 ◽

2012 ◽

Vol 239-240 ◽

pp. 921-931

Author(s):

Jian Zhang ◽

Jun Zhong Zou ◽

Lan Lan Chen ◽

Chen Jie Zhao ◽

Gui Song Wang

Keyword(s):

Time Series ◽

Slow Wave ◽

Digital Signal ◽

Sharp Wave ◽

Epileptiform Discharges ◽

Interictal Epileptiform Discharges ◽

Novel Approach ◽

Recording Process ◽

Eeg Recordings ◽

Maximum Similarity

In this paper, an effective digital signal processing method based on the merger of the increasing and decreasing time-series sequences (MIDS) is introduced. On the basis of the merging of EEG signals, a new IED (Interictal Epileptiform Discharges) detection method is proposed. The first step of this new method is to establish a database by selecting peaked wave fragments. Then, the similarity between pending test fragment and peaked wave samples in the database is calculated. When the maximum similarity is greater than a certain threshold, the fragment is judged to be a peaked wave. Finally, the wave type i.e. spike wave, sharp wave, spike-and-slow wave or sharp-and-slow wave can be determined by whether there is a subsequent slow wave or not. Continuous sharp wave can be viewed as spike rhythm. In this research, 92 IED fragments from 4 suspected epilepsy patients are collected to establish the sample database. The proposed method was tested on EEG recordings from other 31 suspected patients. The results show that 98.11% of the IED fragments marked by doctors were detected. The experimental results show that this method performs well at IED detection in the clinical EEG data. The similarity is measured based on the comparison between fragments of different time length and can be viewed as a novel approach for the detection of typical EEG waveform. This research draws two conclusions: (1) the waveform of individual peaked wave is stable during 24-hour EEG recording process; (2) the database containing a small number peaked wave samples can be used to detect IED fragments.

Download Full-text

Predictive Value of Interictal Epileptiform Discharges During Non-REM Sleep on Scalp EEG Recordings for the Lateralization of Epileptogenesis

Epilepsia ◽

10.1111/j.1528-1157.1998.tb01431.x ◽

1998 ◽

Vol 39 (6) ◽

pp. 628-632 ◽

Cited By ~ 57

Author(s):

Naoto Adachi ◽

Gonzalo Alarcon ◽

Colin D. Binnie ◽

Robert D. C. Elwes ◽

Charles E. Polkey ◽

...

Keyword(s):

Rem Sleep ◽

Predictive Value ◽

Scalp Eeg ◽

Epileptiform Discharges ◽

Interictal Epileptiform Discharges ◽

Eeg Recordings

Download Full-text

Latency of interictal epileptiform discharges in long-term EEG recordings in epilepsy patients

Seizure ◽

10.1016/j.seizure.2015.03.012 ◽

2015 ◽

Vol 29 ◽

pp. 20-25 ◽

Cited By ~ 9

Author(s):

Konrad J. Werhahn ◽

Elisabeth Hartl ◽

Kristin Hamann ◽

Markus Breimhorst ◽

Soheyl Noachtar

Keyword(s):

Epileptiform Discharges ◽

Interictal Epileptiform Discharges ◽

Eeg Recordings

Download Full-text

How Long Should Routine EEG Be Recorded to Get Relevant Information?

Clinical EEG and Neuroscience ◽

10.1177/1550059417698549 ◽

2017 ◽

Vol 49 (5) ◽

pp. 335-341

Author(s):

Hannah Doudoux ◽

Kristina Skaare ◽

Thomas Geay ◽

Philippe Kahane ◽

Jean L. Bosson ◽

...

Keyword(s):

Cognitive Impairment ◽

Relevant Information ◽

The Other ◽

Neurologic Manifestation ◽

Epileptiform Discharges ◽

Interictal Epileptiform Discharges ◽

Eeg Abnormalities ◽

Eeg Recordings ◽

Time Required ◽

Over Time

Objective. The optimal duration of routine EEG (rEEG) has not been determined on a clinical basis. This study aims to determine the time required to obtain relevant information during rEEG with respect to the clinical request. Method. All rEEGs performed over 3 months in unselected patients older than 14 years in an academic hospital were analyzed retrospectively. The latency required to obtain relevant information was determined for each rEEG by 2 independent readers blinded to the clinical data. EEG final diagnoses and latencies were analyzed with respect to the main clinical requests: subacute cognitive impairment, spells, transient focal neurologic manifestation or patients referred by epileptologists. Results. From 430 rEEGs performed in the targeted period, 364 were analyzed: 92% of the pathological rEEGs were provided within the first 10 minutes of recording. Slowing background activity was diagnosed from the beginning, whereas interictal epileptiform discharges were recorded over time. Moreover, the time elapsed to demonstrate a pattern differed significantly in the clinical groups: in patients with subacute cognitive impairment, EEG abnormalities appeared within the first 10 minutes, whereas in the other groups, data could be provided over time. Conclusion. Patients with subacute cognitive impairment differed from those in the other groups significantly in the elapsed time required to obtain relevant information during rEEG, suggesting that 10-minute EEG recordings could be sufficient, arguing in favor of individualized rEEG. However, this conclusion does not apply to intensive care unit patients.

Download Full-text